IBM Watsonx API wrapper package for calling text generation and embedding requests

---

Setup a local environment or install on current:

PyPi package link: WatsonxConnector

pip install WatsonxConnector

---

Usage / Examples

from WatsonxConnector.connector import WatsonxConnector

example_connector_obj = WatsonxConnector(base_url=cp4d_url,
                                         user_name=user_name,
                                         api_key=api_key,
                                         model_id=model_id,
                                         project_id=project_id
                                         )

for model, model_type in example_connector_obj.get_available_models().items():
    print(f"{model}: {model_type}")

example_connector_obj.set_model_params(temperature=0.7, max_new_tokens=800)
print(example_connector_obj.get_model_params())

print(example_connector_obj.generate_text("Write an example python unittest"))

example_connector_obj.set_model_id(model_id="ibm/slate-125m-english-rtrvr")
for embed in example_connector_obj.generate_embedding(["something", "something else"]):
    print(embed)

Example output from above:

codellama/codellama-34b-instruct-hf: text_generation
ibm/granite-13b-chat-v2: text_generation
ibm/slate-125m-english-rtrvr: embedding
ibm/slate-30m-english-rtrvr: embedding
meta-llama/llama-3-70b-instruct: text_generation

{'decoding_method': 'sample', 'max_new_tokens': 800, 'temperature': 0.7, 'top_k': 20, 'top_p': 0.9, 'repetition_penalty': 1.1}

import unittest
from mymodule import my_function

class MyTest(unittest.TestCase):

    def test_my_function(self):
        result = my_function('hello world')
        self.assertEqual(result, 'Hello World')

if __name__ == '__main__':
    unittest.main()

This is an example of a simple Python unit test using the `unittest` module. 
In this case, we define a new class ...

[-0.024465775, -0.0329509, -0.00046468992, -0.017108368, ...  0.017113319, 0.01991029, -0.010878863, 0.005576967]
[-0.037240397, -0.04046932, -0.03115683, -0.034350585, ... 0.020475477, 0.05027425, 0.006514402, 0.0150749]

---

Documentation:

• Extended Reading
• WatsonxConnector Class Breakdown

List of available class methods:

• set_system_prompt()
• set_model_id()
• set_model_params()
• set_api_version()
• set_project_id()
• get_model_id()
• get_auth_token()
• get_model_params()
• get_available_models()
• get_params()
• check_model_type()
• generate_text()
• generate_embedding()
• generate_auth_token()

License

Apache-2.0